Summary of Shielded Deep Reinforcement Learning For Complex Spacecraft Tasking, by Robert Reed et al.
Shielded Deep Reinforcement Learning for Complex Spacecraft Tasking
by Robert Reed, Hanspeter Schaub, Morteza Lahijanian
First submitted to arxiv on: 8 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Autonomous spacecraft control via Shielded Deep Reinforcement Learning (SDRL) is a rapidly growing research area, but it lacks formalized tasking and safety requirements. This paper addresses this issue by using Linear Temporal Logic (LTL) to formalize tasks and safety requirements. The authors propose an approach to construct a reward function from a co-safe LTL specification for effective training in the SDRL framework. Additionally, they investigate methods for constructing shields from safe LTL specifications for spacecraft applications, proposing three designs that provide probabilistic guarantees. Experiments demonstrate how these shields interact with different policies and the flexibility of the reward structure. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Spacecraft control is getting smarter thanks to a new way of learning called Shielded Deep Reinforcement Learning (SDRL). Right now, it’s hard to know what tasks spacecraft should be doing or whether they’re safe. This paper helps solve this problem by using special math rules to define what tasks are important and how to make sure the spacecraft is safe. They also show three different ways to build a “shield” around the spacecraft that can guarantee its safety with a certain level of probability. |
Keywords
* Artificial intelligence * Probability * Reinforcement learning
